Docling is an open-source document parser from the docling-project, built to get documents ready for generative AI. It reads PDF, DOCX, PPTX, XLSX, HTML and images, with advanced PDF understanding of page layout, reading order and table structure, and OCR for scans.
It exports to Markdown, HTML and JSON, and runs locally, so documents don't leave your machine.
Install with pip install docling (Python 3.10+), then convert a file or URL with docling <source> or use DocumentConverter in Python. MIT licensed.